Distribution of Temperature Tolerance Quantitative Trait Loci in Arctic Charr (Salvelinus alpinus) and Inferred Homologies in Rainbow Trout (Oncorhynchus mykiss)

Corresponding author: Moira M. Ferguson, University of Guelph, 50 Stone Rd. East, Guelph, ON N1G 2W1, Canada., mmfergus{at}uoguelph.ca (E-mail)

Communicating editor: M. NOOR

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We searched for quantitative trait loci (QTL) affecting upper temperature tolerance (UTT) in crosses between the Nauyuk Lake and Fraser River strains of Arctic charr (Salvelinus alpinus) using survival analysis. Two QTL were detected by using two microsatellite markers after correcting for experiment-wide error. A comparative mapping approach localized these two QTL to homologous linkage groups containing UTT QTL in rainbow trout (Oncorhynchus mykiss). Additional marginal associations were detected in several families in regions homologous to those with QTL in rainbow trout. Thus, the genes underlying UTT QTL may antedate the divergence of these two species, which occurred by 16 MYA. The data also indicate that one pair of homeologs (ancestrally duplicated chromosomal segments) have contained QTL in Arctic charr since the evolution of salmonids from a tetraploid ancestor 25–100 MYA. This study represents one of the first examples of comparative QTL mapping in an animal polyploid group and illustrates the fate of QTL after genome duplication and reorganization.

COMPARATIVE mapping of genes is rapidly becoming an efficient method to dissect the genetic basis of quantitative trait variation (PFLIEGER et al. 2001 ; DOGANLAR et al. 2002 ). Although common among agriculturally relevant plants (LAN and PATERSON 2000 ; OARD et al. 2000 ; LAUTER and DOEBLEY 2002 ) and animals (ANDERSSON et al. 1998 ; GEORGES 1998 ; DIEZ-TASCON et al. 2001 ), recent comparative linkage mapping in cervids and their domesticated relatives (SLATE et al. 2002A ) has permitted quantitative trait locus (QTL) studies in a natural population of deer (SLATE et al. 2002B ), highlighting a broader applicability of this approach. However, it is becoming increasingly apparent from these studies that gene duplication is an important force in the evolution of genomes, particularly in flowering plants (SOLTIS and SOLTIS 1999 ; OTTO and WHITTON 2000 ). Comparative linkage mapping studies have provided important insight into the function and divergence of duplicated genes (CRONN et al. 1999 ; SCHRANZ et al. 2002 ; SMALL and WENDEL 2002 ), as shown by the identification of QTL on homeologous chromosomal segments in species of Brassica (AXELSSON et al. 2001 ).

Salmonid fishes like rainbow trout (Oncorhynchus mykiss), Atlantic salmon (Salmo salar), brown trout (Salmo trutta), and Arctic charr (Salvelinus alpinus) represent good models for genomic studies following a duplication event, being derived from a tetraploid ancestor 25–100 MYA (ALLENDORF and THORGAARD 1984 ). In addition, as commercially important species, extensive comparative linkage mapping has been undertaken, revealing both broad homology based on conservation of marker linkages (MAY and JOHNSON 1990 ; SAKAMOTO et al. 2000 ; WORAM 2001 ), and significant divergence among species (WORAM et al. 2003 ). Great inter- and intraspecific karyotypic variation, while maintaining a relatively constant number of chromosome arms (100; PHILLIPS and RAB 2001 ), suggests that genome rearrangement by Robertsonian fission-fusion events played an important role in genome evolution of extant salmonid species since their divergence from a common ancestor 16 MYA (PLEYTE et al. 1992 ; ANDERSSON et al. 1995 ; OAKLEY and PHILLIPS 1999 ).

Temperature tolerance is an important trait from both an economic and an evolutionary perspective in fishes, particularly among cool- and cold-water salmonids. Elevated temperatures may negatively affect fitness components, including parameters of growth, development, and reproduction (JOBLING et al. 1995 ; PANKHURST et al. 1996 ). The polygenic basis of upper temperature tolerance (UTT) has recently been demonstrated in selected lines of rainbow trout by the detection of significant QTL on at least nine linkage groups (JACKSON et al. 1998 ; DANZMANN et al. 1999 ; PERRY et al. 2001 ; our unpublished data). Some of these QTL exhibit unpredictable epistatic effects with the genomic background in which they are expressed (DANZMANN et al. 1999 ), as well as sex-specific effects (PERRY 2001 ), highlighting the complexity of this trait in salmonids.

Arctic charr, which extend into boreal circumpolar regions (SCOTT and CROSSMAN 1985 ), include three putative subspecies in North America representing the Laurentian, Arctic, and Labradorean mitochondrial lineages (BRUNNER et al. 2001 ), which have undergone different thermal selection pressures. We used information on known locations of QTL for UTT in rainbow trout to test for similar effects in homologous chromosomal regions of Arctic charr on the basis of the current genetic maps for these species (SAKAMOTO et al. 2000 ; WORAM et al. 2003 , WORAM et al. 2004 ). Arctic charr have higher numbers of acrocentric chromosomes relative to metacentric chromosomes compared to rainbow trout (HARTLEY 1987 ; PHILLIPS and RAB 2001 ) so the comparison will provide important information on the fate of duplicated genes in the face of significant genome reorganization. We also incorporated knowledge of homeologous (duplicated) chromosome arm relationships in rainbow trout to test for the possible conservation of duplicated QTL effects in Arctic charr. We used survival analysis, a novel approach to QTL detection, which tests the relative survival through time of individuals inheriting either allele from a parent when subjected to a thermal challenge.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strain and family history:
The aquaculture strains used in this study were derived from the Nauyuk Lake (Northwest Territories, Canada) and Fraser River (Labrador, Canada) populations approximately four generations ago. These populations not only are separated by large geographic distances, but also are characterized by differences in life history and thermal selection regimes. Nauyuk Lake fish, found in the Canadian sub-Arctic (68° N; GYSELMAN 1994 ), are considered to be much less temperature tolerant than Fraser River fish (56° N; DEMPSON and GREEN 1984 ). The optimal temperatures for Nauyuk Lake brood stock are up to 3° lower than those for the Labrador strain of charr (TABACHEK 1991 ). Furthermore, Nauyuk Lake and Fraser River fish possess distinct mitochondrial genomes and belong to the Arctic and Labradorean lineages of Arctic charr, respectively (BRUNNER et al. 2001 ).

Arctic charr gametes were collected from adults in spawning condition on October 22 and 27, 1998, at Coldwater Hatcheries (Coldwater, Ontario, Canada). Eggs and milt from each individual were transported on ice to the Hagen Aqualab facilities (University of Guelph, Guelph, Ontario, Canada). Crosses were produced by mixing the gametes of charr derived from Fraser River, Nauyuk Lake, and F₁ hybrids between the two strains, yielding four F₁ and one backcross families (Table 1). Incubation of embryos took place at 4° until exogenous feeding was achieved, at which time progeny were transferred to raceways (1 x 3 m). The water source originated from an aquifer (underground spring), whose temperature fluctuated between 10° and 12°. The families were pooled and selectively genotyped to ascertain their family origins following the thermal challenge trials. All rearing practices and thermal challenge experiments followed the University of Guelph Aqualab standard operating procedures for holding salmonid fishes and the Canadian Council for Animal Care guidelines.

Upper temperature tolerance trials:
Progeny were subjected to upper temperature tolerance trials 13 months post-fertilization. Trials were conducted within a single week beginning at 17:00 to minimize effects of seasonal or diurnal changes in physiology. Furthermore, to ensure maximum control of temperature, a stand-alone tank (closed system) was set up that could be programmed and monitored via computer. Feeding was terminated 4 days prior to the thermal challenge, and a random subset of fish was transferred to the experimental tank the preceding evening.

A pilot trial (lot I), where 100 fish were taken randomly from a tank containing individuals from all families, indicated that these particular charr possessed a higher incipient lethal temperature than that suggested by the literature (22.5°; BAROUDY and ELLIOTT 1994 ). As a result, a modification of traditional thermal challenges was employed (JACKSON et al. 1998 ). Temperature was increased from the ambient temperature to the published incipient lethal temperature (22.5°) over a period of 60 min and then kept constant for 30 min. Subsequently, the temperature was raised by 0.5° every 30 min until the end of the trial, resulting in a stepped profile (Fig 1). Air stones were inserted into the tank to aerate and evenly distribute the heated water. Fish were considered to have died when they lost equilibrium and could not right themselves; at this point they were euthanized with an overdose of clove oil (KEEN et al. 1998 ), placed on ice, and given individual tags indicating their time of death. Thermal profiles and temperature at death were recorded by two probes placed at either end of the experimental tank that collected data every 10 sec (BoxCar Pro 3.5). The trial continued until all fish had succumbed to the thermal challenge. Three such thermal challenges (lots II, III, and IV) were required to test all the fish. Body weight (wet weight to the nearest tenth of a gram) and fork length (in millimeters) were recorded and muscle and branchial tissue were sampled. All tissues were frozen at -20° until genetic analyses could be undertaken.

View larger version (48K): In this window In a new window Download PPT slide   Figure 1. Temperature profile representing the thermal challenge to which lot II Arctic charr were exposed. Water temperature and time from 17:00 are indicated on the y-axis and x-axis, respectively. Fish were held at 10° prior to the trial. The ramping zone represents the 60-min period when the temperature was increased to the theoretical incipient lethal temperature of 22° for Arctic charr (BAROUDY and ELLIOTT 1994 ). Subsequently, temperature was increased every 30 min by 0.5° until the end of the trial. When an individual lost equilibrium and was no longer responsive to external stimuli, its time of "death" was recorded and it was euthanized.

Genetic analysis:
DNA was extracted from 25–50 mg of muscle or branchial tissue using the standard phenol chloroform method (BARDAKCI and SKIBINSKI 1994 ) as well as a QIAGEN DNEasy tissue extraction kit. Microsatellite loci were screened in all parents using PCR to detect polymorphisms within families (Table 2 and Table 3). When marker loci were monomorphic (a single band in all fish) or did not amplify (no scoreable product), the next closest alternative was screened on the basis of its proximity on the linkage group. Loci on additional linkage groups were also analyzed. Linkage groups are designated with the prefix "RT-" when referring to rainbow trout and "AC-" when referring to Arctic charr to differentiate between the linkage maps constructed for the two species (SAKAMOTO et al. 2000 ; NICHOLS et al. 2003 ; WORAM et al. 2004 ). We are adopting the rainbow trout linkage group nomenclature of NICHOLS et al. 2003 because those assignments are based upon a larger number of genetic markers, including known marker positions from the SAKAMOTO et al. 2000 map.

Loci were chosen on the basis of previous knowledge of polymorphism in other Arctic charr families (WORAM et al. 2004 ) and known association with QTL in rainbow trout, so that multiple linkage groups would be represented. In particular, all loci linked to UTT QTL in rainbow trout were screened in Arctic charr first. Specifically, significant QTL for UTT have been found on linkage groups RT-21, formerly designated RT-B (SAKAMOTO et al. 2000 ), RT-14 (formerly RT-D), and RT-6 (formerly RT-S; (JACKSON et al. 1998 ; DANZMANN et al. 1999 ; PERRY et al. 2001 ). We have also detected significant QTL effects on RT-1 (formerly RT-18) and RT-15 (formerly RT-8; our unpublished data), although the latter designation is tentative as it is based upon data derived from only 48 progeny in one of the backcross families. Furthermore, suggestive QTL have been found on RT-2, RT-3, RT-8, RT-10, RT-12, RT-16, RT-24, RT-20, and RT-31. In the above set, linkage groups RT-12 and RT-16 represent homeologous pairs, and the identified UTT QTL on both linkage groups map to similar locations, suggesting a conservation of QTL effect (our unpublished data). In addition, homeologies have been identified between RT-2/9, RT-3/25, RT-14/26, RT-23/24, RT-9/20, RT-9/13, RT-27/31, and RT-10/18 (SAKAMOTO et al. 2000 ; NICHOLS et al. 2003 ). We attempted to screen markers from known homeologs of the three significant QTL regions in rainbow trout [i.e., markers from RT-26 (homeologous to RT-14)]. Unfortunately, the homeologous affinities for RT-6 and RT-21 are unknown. Randomly chosen markers from additional linkage groups in Arctic charr that are homologous to rainbow trout homeologs with QTL (i.e., RT-9, RT-13, and RT-27) were also analyzed.

The following PCR program, with slight locus-specific modifications, was used: an initial denaturation cycle of 5 min at 95°, followed by 35 cycles of 1 min at the locus-specific annealing temperature, 1 min at 72°, 1 min at 95°, and a final extension time of 10–20 min at 72°. All loci used in this study, annealing temperatures, and known repeat sequences are presented in Table 2. Alleles were separated on a 6% polyacrylamide denaturing gel and visualized with a fluorescence imaging system (Hitachi FMBIOII). Fragment size was estimated by adding 2 ml of GeneScan 350 (Tamra) size standard (PE Applied Biosystems) to each of several lanes of the gel.

Statistical analysis:
PROBMAX (DANZMANN 1997 ) was used to confirm the familial identity of the progeny through pedigree analysis (families 12-111, 12-114, 21-114, 30-136, and 27-139) using genotypes of up to 38 loci per family. Due to time constraints, families 12-114 and 21-114 were selectively genotyped, with 15–25% of the most- and least-temperature-tolerant fish chosen for analysis (i.e., the tails of the distribution). Selective genotyping is a powerful method for QTL detection, although it may result in biased estimates of allelic effects (e.g., DARVASI and SOLLER 1992 ).

Normality of temperature tolerance data was tested within each family prior to quantitative trait analysis using a Kolmogorov-Smirnov test, which is appropriate for samples sizes used here (Table 4). In addition, the Pearson product-moment correlation was used to determine whether fork length or body weight was associated with upper temperature tolerance in each family.

Due to the potential variability of the temperature profiles across lots (II, III, and IV) and the uneven representation of families within lots, temperature profiles of individual families across lots were compared using the Welch statistic for unequal sample size and variance. For all families except 12-114, mean time until death ("Time"), cumulative temperature profile from 10° acclimation temperature at time of death ("Area"), and "knockdown" temperature ("Temp") were not significantly different across lots (Table 4). Knockdown temperature is defined as the maximal temperature to which an individual fish can survive before it loses equilibrium. Mean temperatures differed across lots for family 12-114. However, regression of "lot" onto temperature tolerance in each family showed that it contributed a negligible amount to the total variance of the model (no increase or a decrease in R²; data not shown). Therefore, progeny from families tested in different lots were pooled for analysis, and "Time" was used as the response variable.

Progeny of heterozygous parents (Table 3) were tested for the expected 1:1 segregation of alleles using the chi-square goodness-of-fit test statistic. Sequential Bonferroni correction for multiple tests was used to ensure an experiment-wide error rate of P < 0.05 within each family (RICE 1989 ). To compensate for the increased likelihood of generating type II errors when this correction is applied to large number of tests, we considered that a correction based upon the number of linkage groups examined was appropriate. Thus, our experiment-wide a = 0.05 level was defined as 0.05/16 and P 0.003, since we examined markers located on 16 Arctic charr linkage groups.

QTL analysis was performed on the maternal and paternal component separately using survival analysis on "Time" to compare the allele classes. The Kaplan Meier product limit measure was employed because it is a nonparametric (or distribution-free) method (KLEINBAUM 1996 ), and the thermal profiles for each allele class had a nonconstant slope. To account for the change in slope midway through the thermal challenge, a censoring variable was included whereby the first 50% of fish to have died were considered "uncensored" (exponent value of 1) and the last 50% "censored" (exponent value of 0). Thus, the survival rate data are included for all individuals up to the censoring point, while individuals surviving past this point are "alive." This parallels the method by which the critical thermal maximum is calculated. The survival analysis performed in this study therefore compares the following survival function of each allele class at a single locus:

In this equation S (t) represents the survival function, n is the total number of cases, denotes the geometric product across all cases less than or equal to t, j is the individual of interest, t is the time interval, and (j) is a constant that is either 1, if the jth case is uncensored, or 0, if it is censored. This estimate of the survival function is the product limit estimator (STATISTICA FOR WINDOWS 1995; KLEINBAUM 1996 ).

This equation compares the proportion of censored and uncensored individuals in each allele class while taking into account the rate of death ("hazard function"); the P-value is derived using Cox's F-test, a test statistic specific to survival analysis, which is appropriate for limited sample sizes (STATISTICA FOR WINDOWS 1995). We dealt with the potential confounding effect of body weight by regressing time of death onto body weight in the families for which body weight significantly correlated with survivorship (12-111, 21-114; data not shown). Thus, we calculated "body weight-corrected time of death" by the formula + residuals; this corrected measure of time was then used in the survival analysis for these two families. However, it is important to note that the results did not change whether body weight was taken into account or not (data not shown).

Genotypic classes were similarly compared in progeny when both the sire and the dam were heterozygous for the same alleles. This was accomplished by scoring homozygotes as 11 or 22 (depending on whether they inherited small or large alleles) and heterozygotes as 12. Then each pairwise comparison was performed (i.e., 11 vs. 22, 11 vs. 12, and 12 vs. 22).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A single locus (Ssa3NUIG in family 12-111) deviated significantly from Mendelian proportions (²_{adj (0.05, 1)} = 9.82, P < 0.05). Loci for which significant deviations from Mendelian segregation were detected were further tested for conformation to 1:1:1:1 genotypic ratios across both parents and did not deviate significantly from expectations (P > 0.05; data not shown).

Marker-UTT associations were detected using survival analysis (Table 5). Significant associations between marker alleles and UTT were detected for Ssa189NVH on linkage group AC-13 in the dam of 27-139 (P < 0.003), as well as for SsaF43NUIG (AC-26) in the sire of 12-111 (P < 0.001). The localization of QTL on linkage group AC-13 is further supported in 12-111 by the detection of suggestive associations (0.003 < P < 0.05) in alleles derived from the male at Ssa85DU (P < 0.046) and at Ssa185NVH (P < 0.021). Similarly, suggestive associations at One10ASC in 12-111 (sire; P < 0.005) and 12-114 (dam; P < 0.021) corroborate the QTL on linkage group AC-26.

The sex-specific distribution of markers on AC-13 suggests that QTL effects are localized in two different regions (region marked by Ssa189NVH in female 27 and a second marked by Ssa85DU, Ssa185NVH, and OmyPuPuPyDU in male 111). It appears that two different QTL exist on AC-13 because the effect in female 27 is confined to a marker (Ssa189NVH) that is unlinked to the region containing the other three markers. Moreover, the effect in male 111 is strongest in the region marked by Ssa85DU, Ssa185NVH, and OmyPuPuPyDU and less so in the region marked by Ssa189NVH (10 cM distant).

Additional suggestive associations between microsatellite loci and UTT were detected in different families on linkage groups AC-4 (sire effect; SSOSL32/i), AC-9 (sire; Ssa14DU), AC-12 (genotypic data; Ssa119NVH), AC-15 (sire; OmyRGT2/iiTUF), AC-19 (dam; OmyRGT46TUF), AC-20 (dam; OmyRGT4TUF; genotypic data; OmyTRCARR), AC-25 (dam; OmyRGT39TUF), and one unassigned marker (genotypic data; OmyOGT5TUF). The QTL on AC-4 and AC-25 are in homeologous regions. The QTL region on AC-4 marked by SSOL32/i appears homeologous to a QTL region on AC-25 marked by RGT39TUF.

In summary, 1 significant and 3 marginal associations were detected out of a total of 37 independent tests across both parents in family 12-111; 2 marginal QTL in 12-114 and 21-114, out of 29 and 27 tests, respectively; no associations in 29 tests in 30-136; and 1 significant and 4 putative associations from a total of 29 tests in 27-139. This corresponds to 2 significant (1%) and 11 marginal (7%) associations between microsatellite loci and UTT detected in 151 tests across five families, when dam and sire components and genotypic data are considered (see Table 3 and Table 6 for more details).

The data suggest that there may be a heterozygote advantage at OmyTRCARR (Table 5). Heterozygous 120/126 individuals showed greater survival relative to the 120/120 homozygote in half-sib family 12-114. Heterozygotes tend to be more temperature tolerant than their homozygous siblings, as assessed by mean survival time and mortality rate. This phenomenon would not have been detected had both parents not been heterozygous for the same alleles (i.e., QTL analysis could not be performed on the whole data set).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Our study is one of the first to undertake comparative QTL mapping in an animal polyploid group. More importantly, our findings highlight the complexities when the taxa have undergone significant genomic reorganization after the polyploid event and have been subjected to very different evolutionary selection pressures. Salmonid fishes like rainbow trout and Arctic charr have long been accepted as important animal models for chromosomal and genetic divergence following a polyploid event (ALLENDORF and THORGAARD 1984 ; MAY and JOHNSON 1990 ; PHILLIPS and RAB 2001 ) but only recently have studies started to address how this molecular architecture relates to phenotypic expression and evolution (PERRY 2001 ; PERRY et al. 2001 ; O'MALLEY et al. 2003 ; WORAM et al. 2003 ). We have detected two significant QTL and seven suggestive QTL for UTT in Arctic charr. Two of the suggestive QTL are found on homeologous linkage groups, indicating functional conservation across duplicated chromosomes. Moreover, comparative mapping suggests that as many as six of these chromosomal regions also have detectable effects in rainbow trout. Thus, some of the genes underlying temperature tolerance QTL may antedate the divergence of Arctic charr and rainbow trout from a common ancestor 16 MYA (ANDERSSON et al. 1995 ).

Determining the homologies of regions between the species, and thus testing if a QTL effect appears to be conserved at the chromosomal level, is complex because of differences in the composition of the marker sets, the karyotypic divergence between the species (PHILLIPS and RAB 2001 ), and the sex-specific recombination rates in salmonid fishes (SAKAMOTO et al. 2000 ). Reduced recombination in males results in inheritance of entire chromosome segments (except in telomeric regions) and leads to an increased ability to detect QTL but a decreased ability to localize the QTL to a particular segment. In contrast, marker-trait associations are less likely to be detected in females because of large interlocus distances. However, once a potential linkage group has been targeted, significant marker-trait associations in females tend to be more representative of true QTL location.

Rainbow trout have UTT QTL in regions that are homologous to those containing the two significant QTL in Arctic charr (JACKSON et al. 1998 ; DANZMANN et al. 1999 ; PERRY et al. 2001 ; Fig 2). First, the QTL marker on RT-24 (Ssa85DU) is in close proximity to the QTL markers on AC-13 (OmyPuPuPyDU and Ssa185NVH; Fig 2A). However, we cannot determine whether the second QTL region on AC-13 shows similar effects across species because Ssa189NVH has yet to be mapped in rainbow trout. Second, the proximity of the QTL marker on AC-26 (SsaF43NUIG) to Cocl3LAV, which in turn maps proximally to the QTL marker on RT-6 (Ssa20.19NUIG), suggests conservation across species (Fig 2B).

View larger version (223K): In this window In a new window Download PPT slide   Figure 2. Comparative homologies between Arctic charr and rainbow trout chromosomes possessing UTT QTL (a solid box indicates UTT regions with P < 0.01 and a cross-hatched box indicates UTT regions where 0.05 > P > 0.01). Male (M) and female (F) specific maps for each species are indicated following the linkage group designation and preceding the mapping panel used. For Arctic charr, two mapping families were used: Family 2 (Fam2) and Family 3 (Fam3). The markers analyzed in the present study are in boldface type. Similarly, in rainbow trout, two main mapping families, lot 25 and lot 44, were used for linkage map construction. For composite maps (involving combined data from both mapping families), the marker polymorphism sources are as follows: Fam2, ^; Fam3, +; lot 25, ^; lot 44, +. Details for the map construction and marker sources are presented in WORAM et al. 2004 .

Four of the suggestive UTT QTL in Arctic charr may also show homologies in rainbow trout and Arctic charr. First, the marker Ssa14DU (AC-9 and RT-14) is associated with differential thermal challenge survival in both species (Fig 2C). Second, the QTL marker on AC-12 (Ssa119NVH) maps syntenically to Omy77DU in males (WORAM et al. 2004 ; Fig 2D). This region on RT-16 also has a suggestive UTT effect in rainbow trout (JACKSON et al. 1998 ; DANZMANN et al. 1999 ). The third case is not as straightforward because of complexities in linkage homologies between the species. The QTL marker (OmyRGT2/iiTUF) on AC-15 falls in the same cluster of markers as the QTL marker (Omy105DU) on RT-10 (JACKSON et al. 1998 ; DANZMANN et al. 1999 ) in the male, suggesting conservation across species (Fig 2E). However, other markers showing zero recombination to Omy105DU on RT-10 map to other linkage groups in Arctic charr (e.g., One10ASC on AC-26). Thus, we cannot be sure of homology until we resolve the linkage arrangements in both sexes. Fourth, the chromosomal region surrounding TRCARR on RT-20 (our unpublished results) and on AC-20 (QTL marker shows zero recombination with TRCARR in males) both contain UTT QTL (Fig 2F).

The detection of QTL in orthologous regions of Arctic charr and rainbow trout supports the findings from many recent comparative QTL studies. Homologies are detectable across both closely and distantly related species (PATERSON et al. 2000 ; LAHBIB-MANSAIS et al. 2003 ). Interestingly, some analyses suggest a remarkable conservation of genetic architecture in that the domestication of the Solanaceae has involved a limited number of loci in the different species (DOGANLAR et al. 2002 ). Our results also support the recurring theme that either gene duplication at the level of entire genomes (polyploidization) and subsequent gene loss (diploidization; BOWERS et al. 2003 ) or intrachromosomal segmental duplication (LOCKE et al. 2003 ) affects the propensity of conservation in gene order and location.

We have limited evidence for the apparent functional retention of duplicate QTL regions in Arctic charr as only one pair of ancestral homeologs had detectable QTL. Marginal evidence that three pairs of ancestral homeologs contained detectable QTL for either spawning date or body weight has been found in rainbow trout out of eight homeologs tested (O'MALLEY et al. 2003 ). In one of the three pairs of rainbow trout homeologs, the duplicated QTL regions mapped to the same relative chromosomal location, while the exact localization of the QTL position in one of the other pairs was difficult to infer since it was based upon data from a male-derived map. In addition, the mapping of body weight QTL to four pairs of homeologous segments in Atlantic salmon (D. REID, A. SZANTO, B. GLEBE, R. DANZMANN and M. FERGUSON, unpublished observations) provides some evidence that ancestrally duplicated chromosomes may retain similar gene function.

The importance of gene duplication and polyploidy in the evolution of phenotypic diversity is more readily apparent when considering plants, of which a large proportion are thought to be polyploid in origin (SOLTIS and SOLTIS 1999 ; OTTO and WHITTON 2000 ). Studies have found a high degree of observed duplication of QTL in autopolyploids (e.g., sugarcane; MING et al. 2002 ) and possibly less so in allopolyploids (e.g., cotton; CRONN et al. 1999 ). The loss or divergence of gene function may also relate to the length of the diploidization process (APARICIO 2000 ). Unfortunately, there are few animal systems in which to test these ideas. Nevertheless, the salmonids are expected to show more divergence of gene function (less conservation of QTL across homeologs) compared to sugarcane where the polyploid event is thought to have occurred within the last few million years (PATERSON et al. 2000 ). The existing salmonid QTL data certainly support this prediction. However, allozyme studies suggest that salmonids exhibit as much as 50% retention of duplicate allozyme expression (ALLENDORF and THORGAARD 1984 ) and thus some conservation of QTL across ancestral homeologs is expected. High rates in the frequency of duplicate preservation have been attributed to relaxed selection or accelerated evolution at replacement sites early on, followed by a gradual increase in selective constraints (LYNCH and CONERY 2000 ), with genes involved in the adaptive environmental and stress responses being significantly overrepresented (KONDRASHOV et al. 2002). Thus, we might expect to see greater rates of QTL conservation across homeologs for certain traits, but not others, once the data become available.

The observation that multiple QTL were detected in pure strain parents (Fraser River and Nauyuk Lake) was unexpected. It was predicted that greater effects would have been detected in the male F₁ hybrid parent due to segregation of QTL alleles, under the assumption that pure strains were almost fixed for alternate alleles. This was inferred because these strains are descended from populations that are adapted to very different thermal regimes as mentioned previously. While the majority of QTL effects were detected in the F₁ male parent, QTL effects were also detected in all the other parents, with the exception of the female parent in one family. Such cryptic variation for temperature tolerance within "pure" strains may have been uncovered upon disruption of the genetic background, as suggested for various invariant phenotypic characters in teosinte when crossed to maize (LAUTER and DOEBLEY 2002 ). The observation that generally more effects were found for the alleles inherited from Fraser strain parents than for those from Nauyuk Lake fish may relate to the numbers of founding individuals used initially to produce the different strains, such that there was more genetic variation in the Fraser than in the Nauyuk Lake fish (LUNDRIGAN 2001 ).

The genetic basis of UTT QTL is not presently known. Evidence in Fugu and Ictalurus indicates that many microsatellites are present in untranscribed regions of genes and even in coding regions (EDWARDS et al. 1998 ; LIU et al. 1999 ), suggesting that microsatellite-based QTL studies across species may be informative with respect to detecting close linkages to functional genes. Although a variety of candidate genes for stress response are known in vertebrates, very few of these have been mapped in salmonid genomes. For example, the only heat-shock cognate mapped in salmonids is localized to linkage group RT-9 in rainbow trout (SAKAMOTO et al. 2000 ), which is homologous to the QTL-containing region on AC-20 in four families of Arctic charr. The exact localization of these markers to the putative candidate gene is difficult to ascertain since the hsc71 gene in rainbow trout was mapped in the male. The expression of heat-shock protein (hsp) families is directly related to thermotolerance (COLEMAN et al. 1995 ), and thus hsp may be primary candidate genes for UTT QTL on AC-20. Future research on candidate genes will provide important insights into the genetic basis of QTL affecting traits of evolutionary interest.

	ACKNOWLEDGMENTS

We thank R. Woram, J. Stout, C. Rexroad, T. Sakamoto, A. Ozaki, N. Okamoto, C. McGowan, W. Davidson, K. Gharbi, R. Guyomard, B. Hoyheim, J. Taggart, R. Powell, and L.-E. Holm for sharing unpublished data on the Arctic charr and rainbow trout linkage maps. This work would not have been possible without the generous donation of gametes by Coldwater Hatcheries (Coldwater, Ontario, Canada) and Alma Research Station (Alma, Ontario, Canada) or without the housing of fish at the Hagen Aqualab. Funding was provided by a Strategic Projects Grant awarded from Natural Sciences and Engineering Research Council of Canada.

Manuscript received March 14, 2003; Accepted for publication July 11, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLENDORF, F. W., and G. H. THORGAARD, 1984 Tetraploidy and the evolution of salmonid fishes, pp. 1–53 in Evolutionary Genetics of Fishes, edited by B. J. TURNER. Plenum Press, New York.

ANDERSSON, E., B. PEIXOTO, V. TORMANEN, and T. MATSUNAGA, 1995  Evolution of the immunoglobin M constant region genes of salmonid fish, rainbow trout (Oncorhynchus mykiss) and Arctic charr (Salvelinus alpinus): implications concerning divergence time of species. Immunogenetics 41:312-315.[Medline]

ANDERSSON, K., L. ANDERSSEN-EKLUND, I. EDFORS-LILJA, H. ELLEGREN, C. HALEY et al., 1998 Case history in animal improvement: genetic mapping of QTLs for growth and fatness in pigs, pp. 241–253 in Molecular Dissection of Complex Traits, edited by A. PATERSON. CRC Press, Cleveland/Boca Raton, FL.

ANGERS, B., L. BERNATCHEZ, A. ANGERS, and L. DESGROSEILLERS, 1995  Specific microsatellite loci for brook charr reveal strong population subdivision on a microgeographic scale. J. Fish Biol. 47(Suppl. A):177-185.

APARICIO, S., 2000  Vertebrate evolution: recent perspectives from fish. Trends Genet. 16:54-56.[Medline]

AXELSSON, T., O. SHAVORSKAYA, and U. LAGERCRANTZ, 2001  Multiple flowering time QTLs within several Brassica species could be the result of duplicated copies of one ancestral gene. Genome 44:856-864.[Medline]

BARDAKCI, F. and D. O. F. SKIBINSKI, 1994  Applications of the RAPD technique in tilapia fish: species and subspecies identification. Heredity 73:117-123.

BAROUDY, E. and J. ELLIOTT, 1994  The critical thermal limits for juvenile Arctic charr Salvelinus alpinus.. J. Fish Biol. 45:1041-1053.

BOWERS, J. E., B. A. CHAPMAN, J. RONG, and A. H. PATERSON, 2003  Unravelling angiosperm genome evolution by phylogenetic analysis of chromosomal duplication events. Nature 422:433-438.[Medline]

BRUNNER, P. C., M. R. DOUGLAS, A. OSINOV, C. C. WILSON, and L. BERNATCHEZ, 2001  Holarctic phylogeography of Arctic charr (Salvelinus alpinus L.) inferred from mitochondrial DNA sequences. Evolution 55:573-586.[Medline]

COLEMAN, J., S. HECKATHORN, and R. HALLBERG, 1995  Heat-shock proteins and thermotolerance: linking molecular and ecological perspectives. Trends Ecol. Evol. 10:305-306.

CRONN, R. C., R. L. SMALL, and J. F. WENDEL, 1999  Duplicated genes evolve independently after polyploid formation in cotton. Proc. Natl. Acad. Sci. USA 96:14406-14411.[Abstract/Free Full Text]

DANZMANN, R. G., 1997  PROBMAX: a computer program for assigning unknown parentage in pedigree analysis from known genotypic pools of parents and progeny. J. Hered. 88:333.[Free Full Text]

DANZMANN, R. G., T. R. JACKSON, and M. M. FERGUSON, 1999  Epistasis in allelic expression at upper temperature tolerance QTL in rainbow trout. Aquaculture 173:45-58.

DARVASI, A. and M. SOLLER, 1992  Selective genotyping for determination of linkage between a marker locus and a quantitative trait locus. Theor. Appl. Genet. 85:353-359.

DEMPSON, J. B. and J. M. GREEN, 1984  Life history of anadromous Arctic charr, Salvelinus alpinus, in the Fraser River, northern Labrador. Can. J. Zool. 63:315-324.

DIEZ-TASCON, C., Y. BASCON, J. J. ARRANZ, F. DE LA FUENTA, and F. SAN PRIMITIVO, 2001  Mapping quantitative trait loci for milk production traits on ovine chromosome 6. J. Dairy Res. 68:389-397.[Medline]

DOGANLAR, S., A. FRARY, M. C. DAUNAY, R. N. LESTER, and S. D. TANKSELY, 2002  Conservation of gene function in the Solanaceae as revealed by comparative mapping of domestication traits in eggplant. Genetics 161:1713-1726.[Abstract/Free Full Text]

EDWARDS, Y. J. K., G. ELGAR, M. S. CLARK, and M. J. BISHOP, 1998  The identification and characterization of microsatellites in the compact genome of the Japanese pufferfish, Fugu rubripes: perspectives in functional and comparative genomic analyses. J. Mol. Biol. 278:843-854.[Medline]

GEORGES, M., 1998 Case history in animal improvement: mapping complex traits in ruminants, pp. 229–239 in Molecular Dissection of Complex Traits, edited by A. PATERSON. CRC Press, Cleveland/Boca Raton, FL.

GYSELMAN, E. C., 1994  Fidelity of anadromous Arctic charr (Salvelinus alpinus) to Nauyuk Lake, N. W. T. Canada. Can. J. Fish. Aquat. Sci. 51:1927-1934.

HARTLEY, S., 1987  The chromosomes of salmonid fishes. Biol. Rev. Camb. Philos. Soc. 62:197-214.

JACKSON, T., M. M. FERGUSON, R. G. DANZMANN, A. FISHBACK, and P. IHSSEN et al., 1998  Identification of two QTL influencing upper temperature tolerance in three rainbow trout (Oncorhynchus mykiss) half-sib families. Heredity 80:143-151.

JOBLING, M., H. JOHNSEN, G. PETTERSEN, and R. HENDERSON, 1995  Effect of temperature on reproductive development in Arctic charr, Salvelinus alpinus (L.). J. Therm. Biol. 20:157-165.

KEEN, J. L., D. L. G. NOAKES, R. D. MOCCIA, and C. G. SOTO, 1998  The efficacy of clove oil as an anaesthetic for rainbow trout, Oncorhynchus mykiss (Walbaum). Aquaculture Res. 29:89-101.

KLEINBAUM, D. G., 1996 Survival Analysis: A Self-Learning Text (Statistics in the Health Sciences). Springer-Verlag, New York.

KRONDRASHOV, F. A., I. B. ROGOZIN, Y. I. WOLF and E. V. KOONIN, 2002 Selection in the evolution of gene duplications. Genome Biol. 3: 0008.1–0008.9.

LAHBIB-MANSAIS, Y., G. TOSSER-KLOPP, S. LEROUX, U. CABAU, and E. KARSENTY et al., 2003  Contribution to high-resolution mapping in pigs with 101 type I markers and progress in comparative map between humans and pigs. Mamm. Genome 14:275-288.[Medline]

LAN, T. H. and A. H. PATERSON, 2000  Comparative mapping of quantitative trait loci sculpting the curd of Brassica oleracea.. Genetics 155:1927-1954.[Abstract/Free Full Text]

LAUTER, N. and J. DOEBLEY, 2002  Genetic variation for phenotypically invariant traits detected in teosinte: implications for the evolution of novel forms. Genetics 160:333-342.[Abstract/Free Full Text]

LIU, Z., G. TAN, P. LI, and R. A. DUNHAM, 1999  Transcribed dinucleotide microsatellites and their associated genes from channel catfish Ictalurus punctatus.. Biochem. Biophys. Res. Comm. 259:190-194.[Medline]

LOCKE, D. P., R. SEAGRAVES, L. CARBONE, N. ARCHIDIACONO, and D. G. ALBERTSON et al., 2003  Large-scale variation among human and great ape genomes determined by array comparative genomic hybridisation. Genome Res. 13:347-357.[Abstract/Free Full Text]

LUNDRIGAN, T. A., 2001 Genetic variation in cultured and natural populations of Arctic charr (Salvelinus alpinus) from North America. M.Sc. Thesis, University of Guelph, Guelph, ON, Canada.

LYNCH, M. and J. S. CONERY, 2000  The evolutionary fate and consequences of duplicated genes. Science 290:1151-1155.[Abstract/Free Full Text]

MAY, B., and K. JOHNSON, 1990 Composite linkage map of salmonid fishes (Salvelinus, Salmo, Oncorhynchus), pp. 151–159 in Genetic Maps: Locus Maps of Complex Genomes, edited by S. J. O'BRIEN. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MING, R., Y. W. WANG, X. DRAYE, P. H. MOORE, and J. E. IRVINE et al., 2002  Molecular dissection of complex traits in autopolyploids: mapping of QTLs affecting sugar yield and related traits in sugarcane. Theor. Appl. Genet. 105:332-345.[Medline]

MORRIS, D. B., K. R. RICHARD, and J. M. WRIGHT, 1996  Microsatellites from rainbow trout (Oncorhynchus mykiss) and their use for genetic study of salmonids. Can. J. Fish. Aquat. Sci. 53:120-126.

NAISH, K. A. and L. K. PARK, 2002  Linkage relationships for 35 new microsatellite loci in chinook salmon Oncorhynchus tshawytscha.. Anim. Genet. 33:312-327.

NICHOLS, K. M., W. P. YOUNG, R. G. DANZMANN, C. REXROAD, and M. NOAKES et al., 2003  An updated genetic linkage map for rainbow trout (Oncorhynchus mykiss). Anim. Genet. 34:102-115.[Medline]

OAKLEY, T. H. and R. B. PHILLIPS, 1999  Phylogeny of salmonine fishes based on growth hormone introns: Atlantic (Salmo) and Pacific (Oncorhynchus) salmon are not sister taxa. Mol. Phylogenet. Evol. 11:381-393.[Medline]

OARD, J., M. COHN, S. LINSCOMBE, D. GEALY, and K. GRAVOIS, 2000  Field evaluation of seed production, shattering, and dormancy in hybrid populations of transgenic rice (Oryza sativa) and the weed, red rice (Oryza sativa). Plant Sci. 157:13-22.[Medline]

O'MALLEY, K. G., T. SAKAMOTO, R. G. DANZMANN, and M. M. FERGUSON, 2003  Quantitative trait loci for spawning date and body weight in rainbow trout: testing for conserved effects across ancestrally duplicated chromosomes. J. Hered. 94:273-284.[Abstract/Free Full Text]

OTTO, S. P. and J. WHITTON, 2000  Polyploid incidence and evolution. Annu. Rev. Genet. 34:401-437.[Medline]

PANKHURST, N., G. PURSER, G. VAN DER KRAAK, P. THOMAS, and G. FORTEATH, 1996  Effect of holding temperature on ovulation, egg fertility, plasma levels of reproductive hormones and in vitro ovarian steroidogenesis in the rainbow trout, Oncorhynchus mykiss.. Aquaculture 146:277-290.

PATERSON, H., J. E. BOWERS, M. D. BUROW, X. DRAYE, and C. G. ELSIK et al., 2000  Comparative genomics of plant chromosomes. Plant Cell 12:1523-1539.[Abstract/Free Full Text]

PERRY, G. M., 2001 The genetic architecture of thermal tolerance and correlated growth-related traits in rainbow trout, Oncorhynchus mykiss. Ph.D. Thesis, University of Guelph, Guelph, ON, Canada.

PERRY, G. M. L., R. G. DANZMANN, M. M. FERGUSON, and J. P. GIBSON, 2001  Quantitative trait loci for upper thermal tolerance in outbred strains of rainbow trout (Oncorhynchus mykiss). Heredity 86:333-341.[Medline]

PFLIEGER, S., V. LEFEBVRE, and M. CAUSSE, 2001  The candidate gene approach in plant genetics: a review. Mol. Breed. 7:275-291.

PHILLIPS, R. B. and P. RAB, 2001  Chromosome evolution in the Salmonidae (Pisces): an update. Biol. Rev. 76:1-25.[Medline]

PLEYTE, K. A., S. D. DUNCAN, and R. B. PHILLIPS, 1992  Evolutionary relationships of the salmonid fish genus Salvelinus inferred from DNA sequences of the first internal transcribed spacer (ITS 1) of ribosomal DNA. Mol. Phylogenet. Evol. 1:223-230.[Medline]

RICE, W. R., 1989  Analyzing tables of statistical tests. Evolution 43:223-225.

SAKAMOTO, T., N. OKAMOTO, and Y. IKEDA, 1996  Application of PCR primer pairs from rainbow trout to detect polymorphisms of CA repeat DNA loci in five confamilial species. Fish. Sci. 62:552-555.

SAKAMOTO, T., R. G. DANZMANN, K. GHARBI, P. HOWARD, and A. OZAKI et al., 2000  A microsatellite linkage map of rainbow trout (Oncorhynchus mykiss) characterized by large sex-specific differences in recombination rates. Genetics 155:1331-1345.[Abstract/Free Full Text]

SCHRANZ, M. E., P. QUIJADA, S. B. SUNG, L. LUKENS, and R. AMASINO et al., 2002  Characterization and effects of the replicated flowering time gene FLC in Brassica rapa.. Genetics 162:1457-1468.[Abstract/Free Full Text]

SCOTT, W. B., and E. J. CROSSMAN, 1985 Freshwater Fishes of Canada, Bulletin 184. Fisheries Research Board of Canada.

SCRIBNER, K. T., J. R. GUST, and R. L. FIELDS, 1996  Isolation and characterization of novel salmon microsatellite loci: cross-species amplification and population genetic applications. Can. J. Fish. Aquat. Sci. 53:833-841.

SLATE, J., T. C. VAN STIJN, R. M. ANDERSON, K. M. MCEWAN, and N. J. MAQBOOL et al., 2002a  A deer (subfamily Cervinae) genetic linkage map and the evolution of ruminant genomes. Genetics 160:1587-1597.[Abstract/Free Full Text]

SLATE, J., P. M. VISSCHER, S. MACGREGOR, D. STEVENS, and M. L. TATE et al., 2002b  A genome scan for quantitative trait loci in a wild population of red deer (Cervus elaphus). Genetics 162:1863-1873.[Abstract/Free Full Text]

SLETTAN, A., I. OLSAKER, and O. LIE, 1997  Segregation studies and linkage analysis of Atlantic salmon microsatellites using haploid genetics. Heredity 78:620-627.

SMALL, R. L. and J. F. WENDEL, 2002  Differential evolutionary dynamics of duplicated paralogous Adh loci in allotetraploid cotton (Gossypium). Mol. Biol. Evol. 19:597-607.[Abstract/Free Full Text]

SOLTIS, D. E. and P. S. SOLTIS, 1999  Polyploidy: recurrent formation and genome evolution. Trends Ecol. Evol. 14:348-352.[Medline]

STATISTICA FOR WINDOWS, 1995 Statsoft, 1984–1995, release version 5.0. Tulsa, OK.

TABACHEK, J. L., 1991 Comparison of Growth Performance of Rainbow Trout (Oncorhynchus mykiss) and Arctic Charr (Salvelinus alpinus), Department of Fisheries and Oceans, Canada.

TAYLOR, E. B., Z. REDENBACH, A. B. COSTELLO, S. M. POLLARD, and C. J. PACAS, 2001  Nested analysis of genetic diversity in northwestern North American char, Dolly Varden (Salvelinus malma) and bull trout (Salvelinus confluentus). Can. J. Fish. Aquat. Sci. 58:406-420.

WORAM, R. A., 2001 Construction of a genetic linkage map for Arctic charr (Salvelinus alpinus) and comparative mapping with Atlantic salmon, brown trout and rainbow trout. M.Sc. Thesis, University of Guelph, Guelph, ON, Canada.

WORAM, R. A., K. GHARBI, T. SAKAMOTO, B. HOYHEIM, and L.-E. HOLM et al., 2003  Comparative genome analysis of the primary sex-determining locus in salmonid fishes. Genome Res. 13:272-280.[Abstract/Free Full Text]

WORAM, R. A., C. MCGOWAN, J. A. STOUT, K. GHARBI, and M. M. FERGUSON et al., 2004  A genetic linkage map for Arctic charr (Salvelinus alpinus): evidence for higher recombination rates and segregation distortion in hybrid versus pure strain mapping parents. Genome in press.

This article has been cited by other articles:

				M. Sekino and M. Hara Linkage Maps for the Pacific Abalone (Genus Haliotis) Based on Microsatellite DNA Markers Genetics, February 1, 2007; 175(2): 945 - 958. [Abstract] [Full Text] [PDF]